使用GROUP BY和OVER

我试图获得按年份和国家/地区分组的总销售额。这按预期工作：使用GROUP BY和OVER

SELECT Year, Country, SUM([Total Sales]) 
FROM Table 
GROUP BY Country, Year;

然后，我必须比较每个国家对一年总销售额的贡献。我这样做：

SELECT Year, Country, SUM([Total Sales]), 
    SUM([Total Sales]) OVER(PARTITION BY Year) 
FROM Table 
GROUP BY Country, Year;

这给了我这个错误：

Column 'Table.Total Sales' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.

来源

2016-09-19 Federico

您需要嵌套的款项：

SELECT Year, Country, SUM([Total Sales]), 
     SUM(SUM([Total Sales])) OVER (PARTITION BY Year) 
FROM Table 
GROUP BY Country, Year;

第一次看到它时，这个语法有点时髦。但是，窗口函数在,GROUP BY之后被评估为。这说的是总和销售额的总和。。。正是你想要的。

来源

2016-09-19 18:00:49

这是因为你可以用SUM()和GROUP BY并与你的窗口功能的窗口级别查询级别不聚集。这是因为窗口函数在应用了结果集之后发生，包括GROUP BY。在那个时间点，[total sales]列仅在原始表格中作为非聚合字段存在，因此您最终会发现无意义。你可以先聚集在子查询，然后做最重要的是你的窗口：

SELECT Year, Country, Total_Sales, SUM([Total_Sales]) OVER(PARTITION BY Year) 
FROM 
(
    SELECT Year, Country, SUM([Total Sales])  
    FROM Table 
    GROUP BY Country, Year 
) sub

来源

2016-09-19 17:54:51 JNevill

您的第一句话是误导性的，因为您可以在不使用子查询的情况下合并窗口函数和聚合函数。 –

我想你可以不用GROUP BY子句通过以下方式：

SELECT 
    Year, 
    Country, 
    SUM([Total Sales]) OVER(), 
    SUM([Total Sales]) OVER (PARTITION BY Year) 
FROM Table

来源

2016-09-19 17:56:10 Dezz

正如提到的错误，您需要在Group by中添加列当前窗口集合函数，但这会导致聚合错误。

如果你想找到的每个company每year的贡献尝试这样的事情

SELECT year, 
     country, 
     Sum([total sales]), 
     OA.year_total, 
     (Sum([total sales])/NULLIF(OA.year_total,0)) * 100 AS Percent_contribution 
FROM table 
     OUTER apply (SELECT Sum(b.[total sales]) AS year_total 
        FROM table b 
        WHERE a.year = b.year) OA 
GROUP BY country, 
      year, 
      year_total;

Outer Apply，这里将计算出每个year的总和。我们可以用总和发现各公司的贡献

使用JOIN

;WITH cte 
    AS (SELECT b.year, Sum(b.[total sales]) AS year_total 
     FROM table b 
     GROUP BY b.year) 
SELECT year, 
     country, 
     Sum([total sales]), 
     b.year_total, 
     (Sum([total sales])/NULLIF(b.year_total, 0)) * 100 AS Percent_contribution 
FROM table a 
     JOIN cte b 
     ON a.year = b.year 
GROUP BY year, 
      country, 
      b.year_total

来源

2016-09-19 17:59:11

另一种方法尝试

SELECT distinct Year, Country, SUM([Total Sales]) OVER(PARTITION BY Country, Year) sumbycountryyear, SUM([Total Sales]) OVER(PARTITION BY Year) sumbyyear 
FROM Table

来源

2016-09-19 18:00:25 Esperento57

使用GROUP BY和OVER

回答

相关问题